Re: Merging the underline attribute at EOL

[Eli Zaretskii <eliz@gnu.org>]

> Date: Tue, 17 Dec 2019 15:13:45 +0100
> From: Ergus <spacibba@aol.com>
> Cc: emacs-devel@gnu.org
> 
> --- a/src/xdisp.c
> +++ b/src/xdisp.c
> @@ -21596,13 +21596,17 @@ extend_face_to_end_of_line (struct it *it)
>  
>    if (FRAME_WINDOW_P (f)
>        && MATRIX_ROW_DISPLAYS_TEXT_P (it->glyph_row)
> -      && face->box == FACE_NO_BOX
>        && FACE_COLOR_TO_PIXEL (face->background, f) == FRAME_BACKGROUND_PIXEL (f)
>  #ifdef HAVE_WINDOW_SYSTEM
>        && !face->stipple
>  #endif
>        && !it->glyph_row->reversed_p
> -      && !Vdisplay_fill_column_indicator)
> +      && !Vdisplay_fill_column_indicator
> +      && (NILP (face->lface[LFACE_EXTEND_INDEX])
> +          ||(face->box == FACE_NO_BOX
> +	     && face->underline == FACE_NO_UNDERLINE
> +	     && face->overline_p == false
> +	     && face->strike_through_p == false)))
>      return;

I don't think I understand the test for the :extend attribute being
non-nil.  Can you explain why you added it?  In general, faces that we
treat as "base face" when we start face merging don't need to have the
:extend attribute set (see, for example, the tool-bar face), but we
still want to apply them to the space past EOL.  Am I missing
something?  Did you bump into some scenario where this attribute
caused some regression or bad results?

Thanks.